Available via license: CC BY 4.0
Content may be subject to copyright.
water
Article
Combination of River Bank Filtration and
Solar-driven Electro-Chlorination Assuring
Safe Drinking Water Supply for River Bound
Communities in India
Philipp Otter 1, * , Pradyut Malakar 2, Cornelius Sandhu 3, Thomas Grischek 3,
Sudhir Kumar Sharma 4, Prakash Chandra Kimothi 4, Gabriele Nüske 5, Martin Wagner 5,
Alexander Goldmaier 1and Florian Benz 1
1
AUTARCON GmbH, D-34117 Kassel, Germany; goldmaier@autarcon.com (A.G.); benz@autarcon.com (F.B.)
2International Centre for Ecological Engineering, University of Kalyani, Kalyani, West Bengal 741235, India;
pradyutmalakar2@gmail.com
3Division of Water Sciences, University of Applied Sciences Dresden, D-01069 Dresden, Germany;
cornelius.sandhu@htw-dresden.de (C.S.); thomas.grischek@htw-dresden.de (T.G.)
4Uttarakhand State Water Supply and Sewerage Organization, Uttarakhand Jal Sansthan (UJS),
Dehradun 263139, India; sudhirksharma10@yahoo.com (S.K.S.); pckimothi@gmail.com (P.C.K.)
5Technologiezentrum Wasser (TZW) Karlsruhe, D-01326 Dresden, Germany;
gabriele.nueske@tzw.de (G.N.); martin.wagner@tzw.de (M.W.)
*Correspondence: otter@autarcon.com; Tel.: +49-561-5061 868 92
Received: 30 October 2018; Accepted: 2 January 2019; Published: 11 January 2019
Abstract:
The supply of safe drinking water in rural developing areas is still a matter of concern,
especially if surface water, shallow wells, and wells with non-watertight headworks are sources for
drinking water. Continuously changing raw water conditions, flood and extreme rainfall events,
anthropogenic pollution, and lacking electricity supply in developing regions require new and
adapted solutions to treat and render water safe for distribution. This paper presents the findings of
a pilot test conducted in Uttarakhand, India, where a river bank filtration (RBF) well was combined
with a solar-driven and online-monitored electro-chlorination system, treating fecal-contaminated
Ganga River water. While the RBF well provided nearly turbidity- and pathogen-free water as
well as buffered fluctuations in source water qualities, the electro-chlorination system provided
disinfection based on the inline conversion of chloride to hypochlorous acid. The conducted sampling
campaigns provided complete disinfection (>6.7 log) and the adequate supply of residual disinfectant
(0.27
±
0.17 mg/L). The system could be further optimized to local conditions and allows the supply
of microbial-safe water for river bound communities, even during monsoon periods and under the
low natural chloride regimes typical for this region.
Keywords:
electro-chlorination; smart villages; disinfection; river bank filtration; rural water supply,
online monitoring
1. Introduction
The Millennium Development Goal to halve the number of people without access to improved
water sources was achieved in 2015—five years ahead of schedule. By that, 2.6 billion people gained
access to improved water sources. However, there is substantial evidence that improved sources of
drinking water, including piped water, can contain fecal contamination and studies estimate that
1.8–2.0 billion people drink such water [
1
–
4
]. Every year 502,000 deaths are caused by diarrheal diseases
that can be attributed to the consumption of unsafe water [
4
]. Especially rural communities are prone
Water 2019,11, 122; doi:10.3390/w11010122 www.mdpi.com/journal/water
Water 2019,11, 122 2 of 17
to having no access to safe drinking water. Lack of infrastructure, technical expertise, user compliance,
as well as the lack of supply of chemicals and electricity have been identified as reasons for the
failure of rural water treatment and supply systems [
5
]. Point-of-use (PoU) treatment approaches
are often considered as alternatives and have shown to reduce the risk of diarrheal infections by
40% [
6
]. However, the effectiveness of PoU disinfection (including chlorination) depends highly on the
comprehension and willingness of the households to apply the treatment systems correctly, especially
under varying source water conditions [
6
–
8
]. Turbidity impedes the application of chlorine and other
disinfection methods. In that case additional filtration is required, increasing the complexity and costs
for PoU treatment. In the end, the responsibility for safe water supply is passed on to the end user and
the educational and motivational efforts required for establishing a reliable application of PoU may
not pay off.
The here presented combination of river bank filtration (RBF) and solar-driven electro-chlorination
(ECl
2
) could be a feasible option for the decentralized treatment of surface water in river bound
communities. Reported data show that RBF can effectively remove many major water pollutants
and micro-pollutants, including particulates, colloids, algae, pathogens, organic as well as inorganic
compounds, microcystins, and heavy metals [
9
,
10
]. Log reductions for total coliforms of 5.5–6.1 and
for bacteriophages of >4.4 were reported by [
11
,
12
]. Total organic carbon (TOC) removal rates of 60%
are possible [
13
]. Whereas conventional treatment methods, like coagulation-filtration, can reduce
the disinfection by-product (DBP) formation potential by 25% [
13
], the reduction can reach 50%–80%
using RBF, without any waste sludge produced. Furthermore, RBF is able to attenuate temperature
peaks and can provide protection against shock loads. Although inorganic contamination is less likely
found in bank filtrate [
9
], oxygen may be depleted during the passage of the water through the bank.
Under anoxic conditions, iron, manganese, and even arsenic can re-mobilize and enter bank filtration
wells [
14
]. During the planning process of RBF abstraction sites such aspects have to be considered
and recommendations for safe management of RBF sites in India were published [15].
In the US, RBF has received log-credits for pathogen removal and is mainly used for the removal
of suspended solids. The sites are often designed with shorter travel times compared to Europe,
where RBF has been widely applied for more than 130 years to produce drinking water along the
Rhine, Elbe, Danube, and Seine rivers. Furthermore, in developing regions, the interest in RBF is
increasing and the feasibility of its application has been evaluated under different hydrological and
hydrogeological conditions (e.g., in India [16], Egypt [17], and Thailand [18]).
However, the application of RBF wells alone does not assure long-term microbial-safe water.
Despite the cited removal rates, monitoring campaigns and risk assessment studies have repeatedly
shown the presence of total coliforms and Escherichia coli (E. coli) in RBF wells, even at greater
distances (48–190 m) to the river bank [
12
,
19
]. In Haridwar (northern India), where the pilot site for
this project is located, such incidents could be related to the seepage of fecal contamination in the
direct vicinity of the wells [
19
]. Further, recontamination may occur also during distribution and
storage, justifying further disinfection. Here, chlorine, in contrast to, for example, UV-treatment or
ultrafiltration (UF), has a long proven record of rendering water safe during storage and distribution—if
handled correctly [
20
,
21
]. In rural communities; however, chlorination systems have failed for the
same reasons as stated above, as they require constant availability of chemicals, skilled personnel
capable in evaluating the chlorine demand of the water, and strict compliance with existing guidelines.
Furthermore, the application of chlorine compounds is challenged by the formation of DBPs if applied
in unfavorable source water conditions. Even though the risks for microbial contamination usually
exceed the adverse side effects of chlorination [
22
,
23
], guideline values for chlorine dose and inorganic
and organic DBP concentrations exist (Table 1).
The inline-electrolytic production of chlorine (ECl
2
) could pose a feasible alternative towards
the dosing of chlorine. Here, gaseous chlorine is produced directly at the anode of an electrolytic
cell from the chloride dissolved in the water that is to be treated (Equation (1)). The chlorine gas
rapidly dissociates in water to hypochlorous acid, being chemically the same oxidizing agent as in
Water 2019,11, 122 3 of 17
conventional chlorine dosing systems (Equation (2)). The chlorine gas production is accompanied by a
decrease of pH (Equation (3)) and the evolution of hydrogen gas at the cathode (Equation (4)) [24].
Table 1. Selected guideline values concerning the chlorination of drinking water.
Parameter Germany EU WHO India IS 10500
Free Available Chlorine (FAC) [mg/L] 1.2 a/0.1–0.3 b>0.5 0.2/1.0
Bromate [µg/L] 10 10 10 -
Chlorate [µg/L] 200 d250 e700 -
Chlorite [µg/L] 200 250 e700 -
Trihalomethanes (THM) [µg/L] 10 b/50 c100 60–300 -
Bromoform [µg/L] 100
Dibromochloromethane [µg/L] 100
Bromodichloromethane [µg/L] 60
Chloroform [µg/L] 200
a
During treatment.
b
At the end of treatment.
c
Point of use.
d
Valid by the time of pilot test; was reduced to
70 µg/L in December 2017. eAs currently proposed [25].
Anodic reaction chlorine: 2Cl−↔Cl2+2e−(1)
Dissociation of chlorine gas in water: Cl2+2H2O↔HClO +Cl−+H3O+(2)
Anodic reaction oxygen: 2H2O↔O2+4H++4e−(3)
Cathodic reaction: 2 H3O++2e−↔H2+2H2O (4)
To power this process a DC voltage is applied to dimension stable (DSA) titanium electrodes
coated with platinum group metals. Studies have shown that coatings comprising iridium and or
ruthenium oxides (MOX electrodes) produce consistently higher chlorine output compared to platinum
coatings [
26
,
27
]. In comparison to the manifold in literature-described boron-doped diamond (BDD)
electrodes, MOX electrodes are less prone to produce DBPs, especially considering chlorate and
perchlorate [28,29].
To control the production of chlorine, fundamental knowledge about the functional
interrelationship between chloride concentration, current, current density, electrode material,
temperature, and source water quality is required and has become available only very recently [
30
].
Systematical evaluation on the effectiveness of the produced disinfecting agents and the potential
formation of disinfection by-products (DBP) has shown that the application of inline-electrolysis is
comparable to the application of hypochlorous acid [
24
,
27
,
28
,
30
]. However, uncertainty towards the
long-term operability, the effectiveness under very low chloride regimes, and elevated hardness levels
persists [27,28].
For the first time a combination of RBF and solar-driven inline-electrolysis was tested in a long
term trial in northern India. The intention of this combination between natural and engineered
solutions (cNES) was to merge the above-mentioned benefits of RBF for surface water treatment
with the benefits of residual chlorination, eliminating the above-mentioned drawbacks of chemical
dosing. The here presented data summarize the findings of two intensive sampling periods conducted
within a two and a half year pilot trial. The main target was to evaluate the pathogen removal and
residual disinfection capacity. The first eight-month sampling campaign lasted from March–November
and included one monsoon season (July–September). The second sampling campaign lasted for two
weeks and was conducted after system optimization. Further, the formation of DBPs and energy
efficiency of this water treatment approach was evaluated and suggestions for long-term operation
and maintenance requirements were derived.
Water 2019,11, 122 4 of 17
2. Materials and Methods
2.1. Bank Filtration at the Haridwar Site
The test was conducted in Haridwar, India, where 68% of the drinking water supply for the city
is produced by RBF [
15
]. The used large diameter well (IW #18) is situated on Pant Dweep Island,
located between the Upper Ganga Canal and the Ganga River (Figure 1). The distance to the nearest
canal bank is 115m. The siting of the well (IW #18) on an island and the significant natural gradient of
the water table result in a high proportion of bank filtrate in the abstracted water. The water table in
the well varies between 6 and 8.5 m bgl.
Water 2018, 10, x FOR PEER REVIEW 4 of 17
2. Materials and Methods
2.1. Bank Filtration at the Haridwar Site
The test was conducted in Haridwar, India, where 68% of the drinking water supply for the city
is produced by RBF [15]. The used large diameter well (IW #18) is situated on Pant Dweep Island,
located between the Upper Ganga Canal and the Ganga River (Figure 1). The distance to the nearest
canal bank is 115m. The siting of the well (IW #18) on an island and the significant natural gradient
of the water table result in a high proportion of bank filtrate in the abstracted water. The water table
in the well varies between 6 and 8.5 m bgl.
Figure 1. Location of the large diameter well on Pant Dweep Island at Haridwar (after [12]).
Studies conducted at two monitoring wells, starting in 2005, revealed that the bank filtrate
contains dissolved organic carbon (DOC) of less than 1 mg/L under aerobic conditions. Trace metals
were found to be below the Indian Standard IS 10500 (1991) limit [12]. The abstracted water from all
the RBF wells in Haridwar only require disinfection and thus are well suited for the conduction of
the pilot test.
2.2. Inline Electrolysis
The inline electrolytic chlorination unit tested during this trial was originally designed for
surface water filtration and disinfection. The ECl
2
cell stack in this pilot had a total surface area of
600 cm
2
and was operated with a maximum current of 5 A, which resulted in a maximum current
density of 8.5 mA/cm
2
. The cells polarity was inverted every 60 minutes to remove
potentially-forming calcareous deposits from the cathode. At very low chloride concentrations the
chlorine production efficiency may not be sufficient to meet the chlorine demand of the water. In
that case, the station automatically reduces its flow rate. This works well; however, it also reduces
the treatment capacity of the station and thus the economic feasibility. In prior studies conducted
with good source water conditions, 10 mg/L of natural chloride in the water has been identified as
the minimum chloride concentration for flow rates up to 100 L/h [26]. In the here described pilot test
an average treatment capacity of 200 L/h was anticipated and the natural chloride concentration of
the bank filtrate was only 14 ± 2 mg/L. Due to that, the pilot station was equipped with an automated
Figure 1. Location of the large diameter well on Pant Dweep Island at Haridwar (after [12]).
Studies conducted at two monitoring wells, starting in 2005, revealed that the bank filtrate contains
dissolved organic carbon (DOC) of less than 1 mg/L under aerobic conditions. Trace metals were
found to be below the Indian Standard IS 10500 (1991) limit [
12
]. The abstracted water from all the RBF
wells in Haridwar only require disinfection and thus are well suited for the conduction of the pilot test.
2.2. Inline Electrolysis
The inline electrolytic chlorination unit tested during this trial was originally designed for surface
water filtration and disinfection. The ECl
2
cell stack in this pilot had a total surface area of 600 cm
2
and was operated with a maximum current of 5 A, which resulted in a maximum current density of
8.5 mA/cm
2
. The cells polarity was inverted every 60 min to remove potentially-forming calcareous
deposits from the cathode. At very low chloride concentrations the chlorine production efficiency
may not be sufficient to meet the chlorine demand of the water. In that case, the station automatically
reduces its flow rate. This works well; however, it also reduces the treatment capacity of the station and
thus the economic feasibility. In prior studies conducted with good source water conditions, 10 mg/L
of natural chloride in the water has been identified as the minimum chloride concentration for flow
rates up to 100 L/h [
26
]. In the here described pilot test an average treatment capacity of 200 L/h was
anticipated and the natural chloride concentration of the bank filtrate was only 14
±
2 mg/L. Due to
Water 2019,11, 122 5 of 17
that, the pilot station was equipped with an automated NaCl dosing system, which would start to
dose NaCl solution into the feed water tank under the following conditions:
(a)
ORPdrinking water tank <ORPtarget and
(b)
VoltageECl2 cell/CurrentECl2 cell ≥3.
If these requirements were met the system would start dosing NaCl solution until either the
ratio of 3 or the target oxidation reduction potential (ORP) was reached. At a constant cell voltage of
12 V, a current of 4 A would be required to stop dosing and the water would then contain a chloride
concentration of about 50 mg/L. As target ORP a value of 720 mV in the first and 700 mV in the
second pilot phase were set. With that a Free Available Chlorine (FAC) conentration of 0.2–0.5 mg/L
was anticipated.
For the first trial, the treatment system was equipped with a 9-inch pressurized vessel containing
Activated Filtration Media (AFM) to remove potential turbidity still present in the bank filtrate.
The filter was automatically backwashed after three days, independent of the quantity of water that
had passed through the filter. During the second short test phase, a second filter was installed after the
ECl
2
cell to remove calcareous deposits that were released from the cathode after polarity inversion
and had slightly increased the turbidity in the final storage water tank.
A submersible pump lifted the bank filtrate into a 2 m
3
feed tank. The water was then pumped
by an internal system pump through Filter 1 and the electrolytic reactor and into a 1 m
3
final water
storage tank (Figure 2). The chlorine production capacity of the cell was increased by adjusting ECl
2
cell current and the flow rate, which was measured with a GEMÜ 850 flow sensor.
Water 2018, 10, x FOR PEER REVIEW 5 of 17
NaCl dosing system, which would start to dose NaCl solution into the feed water tank under the
following conditions:
(a) ORP
< ORP
and
(b) Voltage Current
⁄≥3.
If these requirements were met the system would start dosing NaCl solution until either the
ratio of 3 or the target oxidation reduction potential (ORP) was reached. At a constant cell voltage of
12 V, a current of 4 A would be required to stop dosing and the water would then contain a chloride
concentration of about 50 mg/L. As target ORP a value of 720 mV in the first and 700 mV in the
second pilot phase were set. With that a Free Available Chlorine (FAC) conentration of 0.2–0.5 mg/L
was anticipated.
For the first trial, the treatment system was equipped with a 9-inch pressurized vessel
containing Activated Filtration Media (AFM) to remove potential turbidity still present in the bank
filtrate. The filter was automatically backwashed after three days, independent of the quantity of
water that had passed through the filter. During the second short test phase, a second filter was
installed after the ECl2 cell to remove calcareous deposits that were released from the cathode after
polarity inversion and had slightly increased the turbidity in the final storage water tank.
A submersible pump lifted the bank filtrate into a 2 m3 feed tank. The water was then pumped
by an internal system pump through Filter 1 and the electrolytic reactor and into a 1 m3 final water
storage tank (Figure 2). The chlorine production capacity of the cell was increased by adjusting ECl2
cell current and the flow rate, which was measured with a GEMÜ 850 flow sensor.
Figure 2. Pilot system setting at well IW #18, Haridwar.
2.3. Sampling, Water Analysis, and Monitoring
For water quality analysis, random samples were taken one to two times per week at five
sampling points (SP0 Ganga River, SP1 bank filtrate, SP2 after AFM filtration, SP3 directly after
inline electrolysis, SP4 in final drinking water storage water tank) (Figure 2). Electric conductivity
(Hach CDC 101), dissolved oxygen (LDO 101), and pH (Hach PHC 101) were measured with a Hach
Multimeter HQ40d (Düsseldorf, Germany). The ORP in SP1 (bank filtrate) and SP4 (drinking water)
was measured directly with the pilot system using a Jumo tecLine Rd electrode (Fulda, Germany).
For parameters shown in Table 2, an Aqualytic AL410 (Dortmund, Germany) handheld photometer
was used. Analysis of immediate parameters and parameters in Table 2 were done on site.
Over flow
Drinking
Water Tank
1.00 0 L
Feed
Tank
2.000 L
Na
Cl
AB
Flow Senso r
Treatment
System
Pump
AFM Filter I
ECl2 Cell
Feed
Pump
Dosing pump
Backwash
SP 4
SP 3
SP 2
SP 1
Cont rol
unit
IW 18
Sampl ing point s
SP0 - Gang a River
SP1 - Bank Filtrate
SP2 - Af ter 1s t A ctivate d Filtrat ion M edia (AF M) Fi lter
SP3 - After ECl2 Cell
SP4 - Drinking Water Tank
(after 2nd AFM Filter)
Solar Energy
Supply
water
meter
AFM Filter II only in 2nd test phase
ORP-
SP0
from
Gang a
Figure 2. Pilot system setting at well IW #18, Haridwar.
2.3. Sampling, Water Analysis, and Monitoring
For water quality analysis, random samples were taken one to two times per week at five sampling
points (SP0 Ganga River, SP1 bank filtrate, SP2 after AFM filtration, SP3 directly after inline electrolysis,
SP4 in final drinking water storage water tank) (Figure 2). Electric conductivity (Hach CDC 101),
dissolved oxygen (LDO 101), and pH (Hach PHC 101) were measured with a Hach Multimeter HQ40d
(Düsseldorf, Germany). The ORP in SP1 (bank filtrate) and SP4 (drinking water) was measured directly
with the pilot system using a Jumo tecLine Rd electrode (Fulda, Germany). For parameters shown
in Table 2, an Aqualytic AL410 (Dortmund, Germany) handheld photometer was used. Analysis of
immediate parameters and parameters in Table 2were done on site.
Water 2019,11, 122 6 of 17
Table 2. Parameters and methods used for analysis with the AL410 photometer.
Parameter Wavelength in nm Method Range
Free Available Chlorine (FAC) 530 100: DPD1 0.01–6
Total Chlorine 530 100: DPD3 0.01–6
NH4-N in mg/L 610 60: Indophenole 0.02–1
Cl−in mg/L * 530 90: Silver nitrate/turbidity 0.5–25
Total Hardness in mg/L CaCO3560 200: Metalphthalein 2–50
* Chloride was additionally determined through titration based on APHA Method 4500-Cl−A.
Pathogens, DBPs, and UV-absorption (UVA, wavelength 254 nm) were analyzed in laboratories of
Uttarakhand Jal Sansthan, the state water supply agency, or at TZW Dresden, Germany. Total coliforms
and E. coli were monitored following DIN EN ISO 9308-2 using Colilert
©
Quanti-Tray
©
from IDEXX
Laboratories, Inc. (Westbrook, CT, USA) with a 24 h incubation time. Samples containing chlorine
were quenched using thiosulfate directly after sampling. Turbidity was measured in Nephelometric
Turbidity Units (NTU) with a Turb 430 IR/T from WTW (Weilheim, Germany) following DIN EN
ISO 7027 (Nephelometric Turbidity Unit). Operational parameters, such as electrolytic cell current,
flow rate, power consumption of the pump, and filtration intervals, were monitored using a system
integrated Supervisory Control and Data Acquisition (SCADA) system. The chlorine demand was
determined based on [
31
] by determining the difference between FAC directly after chlorine production
and after 30 min at SP3. Combined chlorine mainly caused by reaction with nitrogen compounds
was determined as the difference between total chlorine and FAC at SP3 and SP4. UVA-254 was
determined using a Lambda 25 PerkinElmer (Waltham, MA, USA) following DIN 38-404-C3. Inorganic
DBP analysis (chlorate, chlorite, perchlorate, bromate) for the first trial period was done for a duration
of four months following DIN EN ISO 10304-4 and TZW lab method, using an ICS 3000 by Thermo
Fischer Scientific (Waltham, MA, USA) having a detection limit of 1
µ
g/L. In the second short time trial,
random samples were analyzed for Trihalomethanes (THMs) following DIN EN ISO 10301, using a
7890A GC/MS by Agilent Technologies (Santa Clara, CA, USA) with a detection limit of 0.1
µ
g/L.
DBP samples were transported to Germany. In order to reduce the number of samples in the first test
phase, all samples of a sampling week (generally 2–3) were mixed, conserved, and then analyzed.
The specific energy demand per m
3
of drinking water produced was calculated using SCADA data,
summarizing the produced water and the power required for running the system. Here, the energy
consumption of the pump lifting the bank filtrate, the pump pushing the water through the system,
the electrolytic cell, and the power supply for the control and online monitoring units, was evaluated.
2.4. Solar Energy Supply System
Due to the potentially non-existing or unreliable electricity supply in future target regions of the
here tested system, the supply with solar photovoltaic (PV) electricity only was evaluated. Planned
or unplanned electricity shortages are permissible when ECl
2
is applied, as the residual disinfectant
assures safe water conditions during water storage. This is one main advantage compared to alternative
disinfection processes based on, for example, UV radiation, whereby power supply has to be always
guaranteed if water is supplied for 24 h per day. For sizing an adequate solar PV system, different
combinations of the photovoltaic (PV) generator size and battery capacities were subject to a sensitivity
analysis using Homer [
32
]. The established model hereby considered the technical parameters given
in Table 3.
Figure 3a shows the clearness indicies and the global horizontal radiation values at the pilot site.
The acutal solar PV generator installed in Haridwar is shown in Figure 3b.
Following the below presented results of the sensitivity analysis, the solar energy supply system
installed at the pilot site comprised a 900 Wp PV generator and 2
×
96 Ah (C10) valve-regluated
lead-acid (VRLA solar batteries).
Water 2019,11, 122 7 of 17
Table 3. Technical parameters for sensitive analysis conducted with Homer.
Parameter Value
Solar radiation
Average monthly clearness indices for Haridwar in kWh/m2d * See Figure 1
Photovoltaic (PV) panel
Slope, Azimuth 20, 0◦West of South
Nominal operational temperature and temperature coefficient 47 ◦C, −0.5%/◦C
PV module Efficiency 14%
PV generator size considered for sensitivity analysis (24 V) 0.6, 0.8, 0.9, and 1.0 kW
Batteries
Minimum state of charge (SOC) 40%
Battery capacity (C10) considered for sensitivity analysis (24 V) 50, 96, 144 Ah
Load
Load, day-to-day variability, time-step-to-time-step variability 70 W, 5%, 5%
Operational duration ** 22 h per day
Constraints
Maximum annual capacity shortage, operational reserve 1% (88 h/a), 10% hourly load
* These data were obtained from the NASA Langley Research Center Atmospheric Science Data Center Surface
Meteorological and Solar Energy (SSE) web portal, supported by the NASA LaRC POWER Project, and are based
on 30-year average meteorological and solar monthly and annual climatology (January 1984–December 2013)
averages. ** A 24-h per day supply of solar-generated electricty is economically not feasible in regions with distinct
rainy seasons.
Water 2018, 10, x FOR PEER REVIEW 7 of 17
Table 3. Technical parameters for sensitive analysis conducted with Homer.
Parameter Value
Solar radiation
Average monthly clearness indices for Haridwar in kWh/m2d * See Figure 1
Photovoltaic (PV) panel
Slope, Azimuth 20, 0° West of South
Nominal operational temperature and temperature coefficient 47 °C, −0.5%/°C
PV module Efficiency 14%
PV generator size considered for sensitivity analysis (24 V) 0.6, 0.8, 0.9, and 1.0 kW
Batteries
Minimum state of charge (SOC) 40%
Battery capacity (C10) considered for sensitivity analysis (24 V) 50, 96, 144 Ah
Load
Load, day-to-day variability, time-step-to-time-step variability 70 W, 5%, 5%
Operational duration ** 22 h per day
Constraints
Maximum annual capacity shortage, operational reserve 1% (88 h/a), 10% hourly load
* These data were obtained from the NASA Langley Research Center Atmospheric Science Data
Center Surface Meteorological and Solar Energy (SSE) web portal, supported by the NASA LaRC
POWER Project, and are based on 30-year average meteorological and solar monthly and annual
climatology (January 1984–December 2013) averages. ** A 24-h per day supply of solar-generated
electricty is economically not feasible in regions with distinct rainy seasons.
Figure 3a shows the clearness indicies and the global horizontal radiation values at the pilot
site. The acutal solar PV generator installed in Haridwar is shown in Figure 3b.
(a) (b)
Figure 3. Clearness indices and monthly global horizontal radiation values (a); and the installed solar
system (b).
Following the below presented results of the sensitivity analysis, the solar energy supply
system installed at the pilot site comprised a 900 Wp PV generator and 2 × 96 Ah (C10)
valve-regluated lead-acid (VRLA solar batteries).
3. Results and Discussion
3.1. System Operation Haridwar
The evaluated trial phase lasted 244 days, including downtimes of in total 24 days. During most
of the time the station operated without technical problems and allowed continuous sampling.
Reasons for downtimes were, for example, low water levels in the well, infrequent cleaning of PV
modules with subsequent power failure, pump failure. Despite the polarity inversion, the operation
in Haridwar was challenged from time-to-time by sudden growth of calcareous deposits on the
electrolytic cell following elevated levels of hardness in the source water. Those are believed to have
0
2
4
6
8
0
0.2
0.4
0.6
0.8
1
Januar
y
A
p
ril Jul
y
October
Global Horizontal
Radiation [kWh/m²*d]
Clearness Index
Daily Radiation Clearness Index
Figure 3.
Clearness indices and monthly global horizontal radiation values (
a
); and the installed solar
system (b).
3. Results and Discussion
3.1. System Operation Haridwar
The evaluated trial phase lasted 244 days, including downtimes of in total 24 days. During most of
the time the station operated without technical problems and allowed continuous sampling. Reasons
for downtimes were, for example, low water levels in the well, infrequent cleaning of PV modules with
subsequent power failure, pump failure. Despite the polarity inversion, the operation in Haridwar was
challenged from time-to-time by sudden growth of calcareous deposits on the electrolytic cell following
elevated levels of hardness in the source water. Those are believed to have initiated crystallization
at the cell allowing fast build-up of deposits. Deposits needed to be manually removed from the cell
using acid.
3.2. Water Quality Parameters Haridwar
Major water quality parameters of the Ganga River water and bank filtrate of the first sampling
period are summarized in Table 4.
Water 2019,11, 122 8 of 17
Table 4. Water quality of the Ganga River water and bank filtrate.
Parameter Mean ±SD Ganga River Mean ±SD Bank Filtrate n Ganga n Bank Filtrate
Pathogens
Total coliforms (MPN/100 mL) 1.07 ×106±1.89 ×1068.87 ×101±1.20 ×10210 30
E. coli (MPN/100 mL) 2.34 ×104±6.34 ×1041.09 ×101±1.79 ×10113 26
Chemical parameters
Hardness (mg/L) 92 ±39 207 ±40 19 41
Chloride * (titrated) (mg/L) 7 ±3 14 ±2 15 20
Ammonium NH4-N (mg/L) 0.15 ±0.07 0.23 ±0.18 34 41
Physico-chemical parameters
Electrical conductivity (µS/cm) 156 ±17 403 ±31 15 38
pH 7.8 ±0.3 7.5 ±0.1 23 43
Temperature (◦C) 25.5 ±2.6 24.8 ±1.1 21 41
ORP [mV] ND 476 ±58 ND 38
Ultraviolet absorbance (UVA-254) (1/m) 51.6 ±28.1 0.8 ±0.9 14
Total organic carbon (TOC) (mg/L) ND 1.96 ±0.49 ** ND 16
* titrated, ** mix values, MPN-most propable number, ND-not determined.
3.3. Turbidity
Turbidity in the Ganga River averaged to 501
±
243 NTU and was reduced to 0.55
±
0.63 NTU
in the bank filtrate (Figure 4), underlining the role of bank filtration as a barrier for particle ingress.
The AFM filter further reduced the turbidity down to 0.30
±
0.34 NTU, and by that substantially
improved water quality prior to the chlorination. This value could be nearly maintained during the
passage in the electrolysis cell but increased to 0.70
±
1.16 NTU with outliers and 0.40
±
0.38 NTU
without outliers, which were caused by filter-breakthrough. The slight but constant increase of turbidity
was traced back to calcareous deposits released from the electrolytic cell after polarity inversion. In the
second short term trial, a second media filter was installed to remove those deposits.
Water 2018, 10, x FOR PEER REVIEW 9 of 17
Figure 4. Turbidity values along water treatment.
3.4. Disinfectant Production
Figure 5 shows the ORP values of the bank filtrate and the final drinking water as well as the
FAC and total chlorine in the drinking water storage tank.
Figure 5. ORP of bank filtrate (SP1), ECl2 effluent (SP4), and chlorine (FAC and total) in ECl2 effluent
(SP4) (a–c).
The ORP increased from 476 ± 58 mV in the bank filtrate to 720 ± 85 mV in the drinking water
tank. FAC and total chlorine values reached 0.27 ± 0.17 mg/L and 0.30 ± 0.16 mg/L, respectively.
Despite the fact that increased ORP values indicate the presence of chlorine, no direct correlation
between both values could be drawn. As reasons for that, the slow reaction time of the ORP sensors
in combination with a relatively small storage volume for the tested flow rates were identified.
Whenever the ORP sensor indicated a low reading, the system automatically increased the chlorine
production and reduced the flow rate. On the other hand, whenever the ORP sensor signaled a high
reading, the system automatically decreased the chlorine production and increased the flow rate.
Due to the small volume in the drinking water tank, an increase or decrease of chlorine
concentration was not detected quickly enough by the ORP sensor. As a consequence, the chlorine
concentration oscillated while the system tried to maintain the target ORP value. Because of this
oscillation, the chlorine value fell from time-to-time below the minimum target value of 0.2 mg/L
and reached high chlorine values of around 1 mg/L. Disabling the automatic flow rate adaptation or
using a larger drinking water storage tank would compensate for the delay of the ORP sensors in
adjusting to changing chlorine levels, and thus could stabilize the chlorine concentration. The
control mechanism was adapted in the second short term trial by removing the automatic adaptation
of the flow rate, which proved to produce more constant results.
0
200
400
600
800
1,000
1,200
Ganga
River (SP0)
22
Turbidity [NTU]
n
0
1
2
3
4
5
6
BF (SP1) After AFM
(SP2)
After ECl2
(SP3)
DW (SP4)
40 36 36 39
n
b)
a)
0
1
2
3
Total
Cl2 DW
(SP4)
FAC
DW
(SP4)
42 42
Chlorine [mg/L]
0
1
2
3
0
240
480
720
21.3 20.5 19.7 17.9 16.11
Chlorine [mg/L]
ORPl [mV]
Date
ORP Bank Filtrate (SP1) ORP Drinking Water (SP4)
Target ORP (SP4) Total Chlorine Drinking Water (SP4)
FAC Drinking Water (SP4) Minimum guideline value IS
Minimum guideline value TrinkwV
0
240
480
720
ORP
BF
(SP1)
ORP
DW
(SP4)
38 40
ORP [mV]
b) c)
a)
monsoon period
nn
Figure 4. Turbidity values along water treatment.
3.4. Disinfectant Production
Figure 5shows the ORP values of the bank filtrate and the final drinking water as well as the FAC
and total chlorine in the drinking water storage tank.
The ORP increased from 476
±
58 mV in the bank filtrate to 720
±
85 mV in the drinking water
tank. FAC and total chlorine values reached 0.27
±
0.17 mg/L and 0.30
±
0.16 mg/L, respectively.
Despite the fact that increased ORP values indicate the presence of chlorine, no direct correlation
between both values could be drawn. As reasons for that, the slow reaction time of the ORP sensors in
combination with a relatively small storage volume for the tested flow rates were identified. Whenever
the ORP sensor indicated a low reading, the system automatically increased the chlorine production
and reduced the flow rate. On the other hand, whenever the ORP sensor signaled a high reading,
the system automatically decreased the chlorine production and increased the flow rate. Due to the
small volume in the drinking water tank, an increase or decrease of chlorine concentration was not
detected quickly enough by the ORP sensor. As a consequence, the chlorine concentration oscillated
Water 2019,11, 122 9 of 17
while the system tried to maintain the target ORP value. Because of this oscillation, the chlorine value
fell from time-to-time below the minimum target value of 0.2 mg/L and reached high chlorine values
of around 1 mg/L. Disabling the automatic flow rate adaptation or using a larger drinking water
storage tank would compensate for the delay of the ORP sensors in adjusting to changing chlorine
levels, and thus could stabilize the chlorine concentration. The control mechanism was adapted in
the second short term trial by removing the automatic adaptation of the flow rate, which proved to
produce more constant results.
Water 2018, 10, x FOR PEER REVIEW 9 of 17
Figure 4. Turbidity values along water treatment.
3.4. Disinfectant Production
Figure 5 shows the ORP values of the bank filtrate and the final drinking water as well as the
FAC and total chlorine in the drinking water storage tank.
Figure 5. ORP of bank filtrate (SP1), ECl2 effluent (SP4), and chlorine (FAC and total) in ECl2 effluent
(SP4) (a–c).
The ORP increased from 476 ± 58 mV in the bank filtrate to 720 ± 85 mV in the drinking water
tank. FAC and total chlorine values reached 0.27 ± 0.17 mg/L and 0.30 ± 0.16 mg/L, respectively.
Despite the fact that increased ORP values indicate the presence of chlorine, no direct correlation
between both values could be drawn. As reasons for that, the slow reaction time of the ORP sensors
in combination with a relatively small storage volume for the tested flow rates were identified.
Whenever the ORP sensor indicated a low reading, the system automatically increased the chlorine
production and reduced the flow rate. On the other hand, whenever the ORP sensor signaled a high
reading, the system automatically decreased the chlorine production and increased the flow rate.
Due to the small volume in the drinking water tank, an increase or decrease of chlorine
concentration was not detected quickly enough by the ORP sensor. As a consequence, the chlorine
concentration oscillated while the system tried to maintain the target ORP value. Because of this
oscillation, the chlorine value fell from time-to-time below the minimum target value of 0.2 mg/L
and reached high chlorine values of around 1 mg/L. Disabling the automatic flow rate adaptation or
using a larger drinking water storage tank would compensate for the delay of the ORP sensors in
adjusting to changing chlorine levels, and thus could stabilize the chlorine concentration. The
control mechanism was adapted in the second short term trial by removing the automatic adaptation
of the flow rate, which proved to produce more constant results.
0
200
400
600
800
1,000
1,200
Ganga
River (SP0)
22
Turbidity [NTU]
n
0
1
2
3
4
5
6
BF (SP1) After AFM
(SP2)
After ECl2
(SP3)
DW (SP4)
40 36 36 39
n
b)
a)
0
1
2
3
Total
Cl2 DW
(SP4)
FAC
DW
(SP4)
42 42
Chlorine [mg/L]
0
1
2
3
0
240
480
720
21.3 20.5 19.7 17.9 16.11
Chlorine [mg/L]
ORPl [mV]
Date
ORP Bank Filtrate (SP1) ORP Drinking Water (SP4)
Target ORP (SP4) Total Chlorine Drinking Water (SP4)
FAC Drinking Water (SP4) Minimum guideline value IS
Minimum guideline value TrinkwV
0
240
480
720
ORP
BF
(SP1)
ORP
DW
(SP4)
38 40
ORP [mV]
b) c)
a)
monsoon period
nn
Figure 5.
ORP of bank filtrate (SP1), ECl
2
effluent (SP4), and chlorine (FAC and total) in ECl
2
effluent
(SP4) (a–c).
3.5. Pathogens
The analytical results of total coliforms and E. coli, as indicator pathogens, for the first trial period
are shown in Figures 6and 7.
Water 2018, 10, x FOR PEER REVIEW 10 of 17
3.5. Pathogens
The analytical results of total coliforms and E. coli, as indicator pathogens, for the first trial
period are shown in Figure 6; Figure 7.
Figure 6. Total coliforms in the Ganga River water and bank filtrate (a) and drinking water (b), and
statistical interpretation (c); trial period March–November.
Figure 7. E. coli in the Ganga River water and bank filtrate (a) and drinking water (b), and statistical
interpretation (c); trial period March–November.
The bank filtration achieved a log10 reduction of 3.9 and 3.6 for total coliforms and E. coli,
respectively. It is assumed that the peak values in the bank filtrate did not originate from the Ganga
River water, but rather came from seepage into the well from above, as described in [19]. However,
the ECl2 system completely removed still present fecal indicators and water could be kept
microbially safe at all times. The maximum log reduction of the RBF + ECl2 cNES achieved was >6.7
for total coliforms and >5.4 for E. coli. It can be assumed that even higher log reductions could be
reached, considering presence of FAC in the treated water.
0.E+00
5.E+00
1.E+01
21.3 20.5 19.7 17.9 16.11
Date
Treated Water
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
1.E+07
Total Coliforms [MPN/100mL]
Ganga River
Bank Filtrate
monsoon period
a)
b)
Ganga
River (SP0)
Bank
Filtrate
(SP1)
Drinking
Water
(SP4)
10 30 30
c)
n
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
1.E+07
E.Coli [MPN/100mL]
Ganga River
Bank Filtrate
monsoon period
0.E+00
5.E+00
1.E+01
21.3 20.5 19.7 17.9 16.11
Date
Treated Water
b)
Ganga
River (SP0)
Bank
Filtrate
(SP1)
Drinking
Water
(SP4)
13 26 28
a)
c)
n
Figure 6.
Total coliforms in the Ganga River water and bank filtrate (
a
) and drinking water (
b
),
and statistical interpretation (c); trial period March–November.
Water 2019,11, 122 10 of 17
Water 2018, 10, x FOR PEER REVIEW 10 of 17
3.5. Pathogens
The analytical results of total coliforms and E. coli, as indicator pathogens, for the first trial
period are shown in Figure 6; Figure 7.
Figure 6. Total coliforms in the Ganga River water and bank filtrate (a) and drinking water (b), and
statistical interpretation (c); trial period March–November.
Figure 7. E. coli in the Ganga River water and bank filtrate (a) and drinking water (b), and statistical
interpretation (c); trial period March–November.
The bank filtration achieved a log10 reduction of 3.9 and 3.6 for total coliforms and E. coli,
respectively. It is assumed that the peak values in the bank filtrate did not originate from the Ganga
River water, but rather came from seepage into the well from above, as described in [19]. However,
the ECl2 system completely removed still present fecal indicators and water could be kept
microbially safe at all times. The maximum log reduction of the RBF + ECl2 cNES achieved was >6.7
for total coliforms and >5.4 for E. coli. It can be assumed that even higher log reductions could be
reached, considering presence of FAC in the treated water.
0.E+00
5.E+00
1.E+01
21.3 20.5 19.7 17.9 16.11
Date
Treated Water
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
1.E+07
Total Coliforms [MPN/100mL]
Ganga River
Bank Filtrate
monsoon period
a)
b)
Ganga
River (SP0)
Bank
Filtrate
(SP1)
Drinking
Water
(SP4)
10 30 30
c)
n
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
1.E+07
E.Coli [MPN/100mL]
Ganga River
Bank Filtrate
monsoon period
0.E+00
5.E+00
1.E+01
21.3 20.5 19.7 17.9 16.11
Date
Treated Water
b)
Ganga
River (SP0)
Bank
Filtrate
(SP1)
Drinking
Water
(SP4)
13 26 28
a)
c)
n
Figure 7.
E. coli in the Ganga River water and bank filtrate (
a
) and drinking water (
b
), and statistical
interpretation (c); trial period March–November.
The bank filtration achieved a log
10
reduction of 3.9 and 3.6 for total coliforms and E. coli,
respectively. It is assumed that the peak values in the bank filtrate did not originate from the Ganga
River water, but rather came from seepage into the well from above, as described in [
19
]. However,
the ECl
2
system completely removed still present fecal indicators and water could be kept microbially
safe at all times. The maximum log reduction of the RBF + ECl
2
cNES achieved was >6.7 for total
coliforms and >5.4 for E. coli. It can be assumed that even higher log reductions could be reached,
considering presence of FAC in the treated water.
3.6. Chlorine Demand
During the trial period the chlorine demand (
∆
FAC) was very low, with 0.03
±
0.03 mg/L
on average, but peaked to 0.33 mg/L. Combined chlorine, formed directly at SP3 and SP4,
were 0.09 ±0.08 mg/L and 0.02 ±0.02 mg/L, respectively (Figure 8).
Water 2018, 10, x FOR PEER REVIEW 11 of 17
3.6. Chlorine Demand
During the trial period the chlorine demand (ΔFAC) was very low, with 0.03 ± 0.03 mg/L on
average, but peaked to 0.33 mg/L. Combined chlorine, formed directly at SP3 and SP4, were 0.09 ±
0.08 mg/L and 0.02 ± 0.02 mg/L, respectively (Figure 8).
Figure 8. Combined chlorine formed at SP3 and SP4 as well as chlorine demand.
The low average values resulted from the low concentrations of ammonium and organics in the
bank filtrate (see Table 4) and do not indicate any critical potential for organic disinfection
by-product formation. However, as there are substantial fluctuations of organic and nitrogen
compounds in the well water, an automated adaption of the chlorine production process is required
to compensate for changing chlorine demand and combined chlorine formation. Even though there
was no good correlation between ORP and chlorine concentration, the ORP can indicate insufficient
supply of disinfectant and can; therefore, next to cell current and flow rate, be consideredas an
additional parameter to control chlorine production.
3.7. Electrical Conductivity and Chloride Concentration
The correlation between electrical conductivity and chloride concentration is shown in Figure 9.
Figure 9. Correlation between conductivity and chloride concentration in treated water (SP4).
The scatterplot indicates that, especially at higher chloride concentrations, the effect of the
chloride on the conductivity prevails towards other ions. In order to limit the NaCl consumption, the
chloride concentration was supposed to be kept below 50 mg/L. The incidents when higher
concentrations occurred could be tracked back to either a nearly empty feed tank, into which the
dosing pump dosed too much chloride for the water available, or to calcearous deposits that were
formed on the cell. Those deposits have hampered the ability to reach a voltage/current ratio of ≤3.
3.8. Inorganic Disinfection By-Products
The concentrations of the mixed samples for chloride and chlorate in SP3 (directly after the
electrolysis cell) are shown in Figure 10.
0.0
0.2
0.4
Δ FAC (30 min)
SP3
Cl2 comb. SP3 Cl2 comb. SP4
42 42 40
Chlorine [mg/L]
n
R² = 0.9863
250
500
750
1000
1250
1500
0 50 100 150 200 250 300
EC SP4 [µS/cm]
Chloride [mg/L]
Figure 8. Combined chlorine formed at SP3 and SP4 as well as chlorine demand.
The low average values resulted from the low concentrations of ammonium and organics in the
bank filtrate (see Table 4) and do not indicate any critical potential for organic disinfection by-product
formation. However, as there are substantial fluctuations of organic and nitrogen compounds in
the well water, an automated adaption of the chlorine production process is required to compensate
for changing chlorine demand and combined chlorine formation. Even though there was no good
Water 2019,11, 122 11 of 17
correlation between ORP and chlorine concentration, the ORP can indicate insufficient supply of
disinfectant and can; therefore, next to cell current and flow rate, be consideredas an additional
parameter to control chlorine production.
3.7. Electrical Conductivity and Chloride Concentration
The correlation between electrical conductivity and chloride concentration is shown in Figure 9.
Water 2018, 10, x FOR PEER REVIEW 11 of 17
3.6. Chlorine Demand
During the trial period the chlorine demand (ΔFAC) was very low, with 0.03 ± 0.03 mg/L on
average, but peaked to 0.33 mg/L. Combined chlorine, formed directly at SP3 and SP4, were 0.09 ±
0.08 mg/L and 0.02 ± 0.02 mg/L, respectively (Figure 8).
Figure 8. Combined chlorine formed at SP3 and SP4 as well as chlorine demand.
The low average values resulted from the low concentrations of ammonium and organics in the
bank filtrate (see Table 4) and do not indicate any critical potential for organic disinfection
by-product formation. However, as there are substantial fluctuations of organic and nitrogen
compounds in the well water, an automated adaption of the chlorine production process is required
to compensate for changing chlorine demand and combined chlorine formation. Even though there
was no good correlation between ORP and chlorine concentration, the ORP can indicate insufficient
supply of disinfectant and can; therefore, next to cell current and flow rate, be consideredas an
additional parameter to control chlorine production.
3.7. Electrical Conductivity and Chloride Concentration
The correlation between electrical conductivity and chloride concentration is shown in Figure 9.
Figure 9. Correlation between conductivity and chloride concentration in treated water (SP4).
The scatterplot indicates that, especially at higher chloride concentrations, the effect of the
chloride on the conductivity prevails towards other ions. In order to limit the NaCl consumption, the
chloride concentration was supposed to be kept below 50 mg/L. The incidents when higher
concentrations occurred could be tracked back to either a nearly empty feed tank, into which the
dosing pump dosed too much chloride for the water available, or to calcearous deposits that were
formed on the cell. Those deposits have hampered the ability to reach a voltage/current ratio of ≤3.
3.8. Inorganic Disinfection By-Products
The concentrations of the mixed samples for chloride and chlorate in SP3 (directly after the
electrolysis cell) are shown in Figure 10.
0.0
0.2
0.4
Δ FAC (30 min)
SP3
Cl2 comb. SP3 Cl2 comb. SP4
42 42 40
Chlorine [mg/L]
n
R² = 0.9863
250
500
750
1000
1250
1500
0 50 100 150 200 250 300
EC SP4 [µS/cm]
Chloride [mg/L]
Figure 9. Correlation between conductivity and chloride concentration in treated water (SP4).
The scatterplot indicates that, especially at higher chloride concentrations, the effect of the chloride
on the conductivity prevails towards other ions. In order to limit the NaCl consumption, the chloride
concentration was supposed to be kept below 50 mg/L. The incidents when higher concentrations
occurred could be tracked back to either a nearly empty feed tank, into which the dosing pump dosed
too much chloride for the water available, or to calcearous deposits that were formed on the cell.
Those deposits have hampered the ability to reach a voltage/current ratio of ≤3.
3.8. Inorganic Disinfection By-Products
The concentrations of the mixed samples for chloride and chlorate in SP3 (directly after the
electrolysis cell) are shown in Figure 10.
Water 2018, 10, x FOR PEER REVIEW 12 of 17
Figure 10. Chloride and chlorate concentrations during field test (a,b), and their correlation (c).
Even after the long storage period of several weeks until analysis in Germany, the chlorate
concentrations reached only 22 ± 29 µg/L and is not of concern considering WHO and German
guideline values (Table 1). Uncertainty towards the maximum chlorate values exists due to the
mixing of two to three random samples into one sample per week, as the concentration of samples
with higher concentration might have been lowered with samples of lower concentration. However,
the correlation between chloride concentration and chlorate production (Figure 10c) show that
higher chloride concentrations are required to reach elevated levels of chlorate. The two maximum
chlorate concentrations above 100 µg/L went along with an excess of chloride added into the feed
tank. Chlorite and perchlorate were always below the detection limit of 1 µg/L and are; therefore,
not of concern when water is disinfected by means of inline-electrolysis with the here applied
MOX-electrodes.
3.9. Hardness
Figure 11 shows the total hardness values measured during the first system trial.
Figure 11. Total hardness values in SP0 and SP1 during the first long term test (a,b); and calcareous
deposits on the ECl
2
-cell (c).
The hardness values were fluctuating throughout the test period in the Ganga River
(92 ± 39 mg/L) and the bank filtrate (207 ± 40 mg/L). Whereas the values in the bank filtrate ranged
around an unproblematic 200 mg/L during monsoon, the levels reached ~300 mg/L before and after
the monsoon season. Those values have shown to be problematic for system operation, as
spontaneous growth of calcareous deposits on the ECl
2
cell intermittently reduced chlorine
production efficiency and required extra maintenance.
R² = 0.9425
0 50 100 150
Chlorate [µg/L]
Chloride [mg/L]
0
20
40
60
80
100
120
140
21.3 20.5 19.7 17.9 16.11
Chloride [mg/L] and Chlorate [µg/L]
Date
Chlorate after electrolytic cell (SP3mix)
Chloride after electrolytic cell (SP3mix)
Chloride Bank Filtrate (SP1mix) for visibility reduced
in size
monsoon period
Failure
of NaCl
dosing
chloride
SP1mix
chloride
SP3mix
chlorate
SP3mix
16 20 19
a) c)b)
Figure 10. Chloride and chlorate concentrations during field test (a,b), and their correlation (c).
Even after the long storage period of several weeks until analysis in Germany, the chlorate
concentrations reached only 22
±
29
µ
g/L and is not of concern considering WHO and German
guideline values (Table 1). Uncertainty towards the maximum chlorate values exists due to the mixing
of two to three random samples into one sample per week, as the concentration of samples with
higher concentration might have been lowered with samples of lower concentration. However,
the correlation between chloride concentration and chlorate production (Figure 10c) show that
Water 2019,11, 122 12 of 17
higher chloride concentrations are required to reach elevated levels of chlorate. The two maximum
chlorate concentrations above 100
µ
g/L went along with an excess of chloride added into the
feed tank. Chlorite and perchlorate were always below the detection limit of 1
µ
g/L and are;
therefore, not of concern when water is disinfected by means of inline-electrolysis with the here
applied MOX-electrodes.
3.9. Hardness
Figure 11 shows the total hardness values measured during the first system trial.
Water 2018, 10, x FOR PEER REVIEW 12 of 17
Figure 10. Chloride and chlorate concentrations during field test (a,b), and their correlation (c).
Even after the long storage period of several weeks until analysis in Germany, the chlorate
concentrations reached only 22 ± 29 µg/L and is not of concern considering WHO and German
guideline values (Table 1). Uncertainty towards the maximum chlorate values exists due to the
mixing of two to three random samples into one sample per week, as the concentration of samples
with higher concentration might have been lowered with samples of lower concentration. However,
the correlation between chloride concentration and chlorate production (Figure 10c) show that
higher chloride concentrations are required to reach elevated levels of chlorate. The two maximum
chlorate concentrations above 100 µg/L went along with an excess of chloride added into the feed
tank. Chlorite and perchlorate were always below the detection limit of 1 µg/L and are; therefore,
not of concern when water is disinfected by means of inline-electrolysis with the here applied
MOX-electrodes.
3.9. Hardness
Figure 11 shows the total hardness values measured during the first system trial.
Figure 11. Total hardness values in SP0 and SP1 during the first long term test (a,b); and calcareous
deposits on the ECl
2
-cell (c).
The hardness values were fluctuating throughout the test period in the Ganga River
(92 ± 39 mg/L) and the bank filtrate (207 ± 40 mg/L). Whereas the values in the bank filtrate ranged
around an unproblematic 200 mg/L during monsoon, the levels reached ~300 mg/L before and after
the monsoon season. Those values have shown to be problematic for system operation, as
spontaneous growth of calcareous deposits on the ECl
2
cell intermittently reduced chlorine
production efficiency and required extra maintenance.
R² = 0.9425
0 50 1 00 150
Chlorate [µg/L]
Chloride [mg/L]
0
20
40
60
80
100
120
140
21.3 20.5 19.7 17.9 16.11
Chloride [mg/L] and Chlorate [µg/L]
Date
Chlorate after electrolytic cell (SP3mix)
Chloride after electrolytic cell (SP3mix)
Chloride Bank Filtrate (SP1mix) for visibility reduced
in size
monsoon period
Failure
of NaCl
dosing
chloride
SP1mix
chloride
SP3mix
chlorate
SP3mix
16 20 19
a) c)b)
Figure 11.
Total hardness values in SP0 and SP1 during the first long term test (
a
,
b
); and calcareous
deposits on the ECl2-cell (c).
The hardness values were fluctuating throughout the test period in the Ganga River
(92
±
39 mg/L) and the bank filtrate (207
±
40 mg/L). Whereas the values in the bank filtrate ranged
around an unproblematic 200 mg/L during monsoon, the levels reached ~300 mg/L before and after
the monsoon season. Those values have shown to be problematic for system operation, as spontaneous
growth of calcareous deposits on the ECl
2
cell intermittently reduced chlorine production efficiency
and required extra maintenance.
3.10. Second Optimized Test Phase
In the second short term test the automatic flow rate adaption was disabled and constant flow
rates of 160, 220, and 280 L/h were established This was giving the ORP sensor sufficient time to detect
changing chlorine concentrations and allowed the ability to test the system’s reaction on changing
chlorine demand in SP4, adjusting the cell current only. The main results of the second short term test
are presented in Figure 12.
Despite the very short duration of this second pilot trial, it was sufficient to show that the
fluctuation around the set target ORP of 700 mV could be reduced to an acceptable level. The constant
flow rates permitted an adequate utilization of the ORP reading for controlling the chlorine production
and keeping the concentration in the desired range of 0.26
±
0.04 mg/L. The presence of FAC in this
concentration range was represented by elevated ORP values of ~700 mV. It can be assumed that a
larger drinking water storage tank (SP4) would have had a similar effect. Pathogens could still be
completely removed through ECl2.
Further, the effect of the second AFM filter, placed behind the electrolytic cell to remove calcareous
deposits, and the THM concentrations are shown in Figure 13b.
Water 2019,11, 122 13 of 17
Water 2018, 10, x FOR PEER REVIEW 13 of 17
3.10. Second Optimized Test Phase
In the second short term test the automatic flow rate adaption was disabled and constant flow
rates of 160, 220, and 280 L/h were established This was giving the ORP sensor sufficient time to
detect changing chlorine concentrations and allowed the ability to test the system’s reaction on
changing chlorine demand in SP4, adjusting the cell current only. The main results of the second
short term test are presented in Figure 12.
Figure 12. ORP and chlorine concentration in SP4 (a,b); and pathogenic contamination in short term
test with optimized system setting (c).
Despite the very short duration of this second pilot trial, it was sufficient to show that the
fluctuation around the set target ORP of 700 mV could be reduced to an acceptable level. The
constant flow rates permitted an adequate utilization of the ORP reading for controlling the chlorine
production and keeping the concentration in the desired range of 0.26 ± 0.04 mg/L. The presence of
FAC in this concentration range was represented by elevated ORP values of ~700 mV. It can be
assumed that a larger drinking water storage tank (SP4) would have had a similar effect. Pathogens
could still be completely removed through ECl2.
Further, the effect of the second AFM filter, placed behind the electrolytic cell to remove
calcareous deposits, and the THM concentrations are shown in Figure 13b.
Figure 13. Turbidity removal (a,b) and THM formation (c) during short term test with optimized
system setting.
The second filter reduced the turbidity down to 0.55 ± 0.08 NTU, after it had increased to 0.92 ± 0.13
NTU behind the electrolytic cell, improving overall water quality. The THM analysis showed
concentrations of 2.2 ± 0.5 µg/L in SP3 after 30 minutes and 4.2 ± 2.1 µg/L in SP4. Those low
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
1.E+07
Ganga River
(SP0)
Bank Filtrate
(SP1)
Drinking
Water (SP4)
444
Total coliform [MPN/100 mL]
n
a)
0
1
2
3
FAC
SP4
Total
Cl2 SP4
77
b)
n
0
1
2
3
0
350
700
19.7 24.7 29.7 3.8
Chlorine [mg/L]
ORP [mV]
Date
Target ORP (SP4)
ORP Drinking Water (SP4)
FAC Drinking Water (S P4)
Total Chlorine Drinking Water (SP4)
Minimum guideline value IS
c)
0
100
200
300
Ganga River
(SP0)
6
Tur bid ity [N TU]
n
0
2
4
6
8
10
THM (SP3)
[µg/L]
THM (SP4)
[µg/L]
66
THM [µg/L]
n
0
1
2
3
4
5
6
Bank Filtrate
(SP1)
After AFM
(SP2)
After ECl2
(SP3)
Drinking Water
(SP4)
6666
n
b)
a) c)
Figure 12.
ORP and chlorine concentration in SP4 (
a
,
b
); and pathogenic contamination in short term
test with optimized system setting (c).
Water 2018, 10, x FOR PEER REVIEW 13 of 17
3.10. Second Optimized Test Phase
In the second short term test the automatic flow rate adaption was disabled and constant flow
rates of 160, 220, and 280 L/h were established This was giving the ORP sensor sufficient time to
detect changing chlorine concentrations and allowed the ability to test the system’s reaction on
changing chlorine demand in SP4, adjusting the cell current only. The main results of the second
short term test are presented in Figure 12.
Figure 12. ORP and chlorine concentration in SP4 (a,b); and pathogenic contamination in short term
test with optimized system setting (c).
Despite the very short duration of this second pilot trial, it was sufficient to show that the
fluctuation around the set target ORP of 700 mV could be reduced to an acceptable level. The
constant flow rates permitted an adequate utilization of the ORP reading for controlling the chlorine
production and keeping the concentration in the desired range of 0.26 ± 0.04 mg/L. The presence of
FAC in this concentration range was represented by elevated ORP values of ~700 mV. It can be
assumed that a larger drinking water storage tank (SP4) would have had a similar effect. Pathogens
could still be completely removed through ECl2.
Further, the effect of the second AFM filter, placed behind the electrolytic cell to remove
calcareous deposits, and the THM concentrations are shown in Figure 13b.
Figure 13. Turbidity removal (a,b) and THM formation (c) during short term test with optimized
system setting.
The second filter reduced the turbidity down to 0.55 ± 0.08 NTU, after it had increased to 0.92 ± 0.13
NTU behind the electrolytic cell, improving overall water quality. The THM analysis showed
concentrations of 2.2 ± 0.5 µg/L in SP3 after 30 minutes and 4.2 ± 2.1 µg/L in SP4. Those low
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
1.E+07
Ganga River
(SP0)
Bank Filtrate
(SP1)
Drinking
Water (SP4)
444
Total coliform [MPN/100 mL]
n
a)
0
1
2
3
FAC
SP4
Total
Cl2 SP4
77
b)
n
0
1
2
3
0
350
700
19.7 24.7 29.7 3.8
Chlorine [mg/L]
ORP [mV]
Date
Target ORP (SP4)
ORP Drinking Water (SP4)
FAC Drinking Water (S P4)
Total Chlorine Drinking Water (SP4)
Minimum guideline value IS
c)
0
100
200
300
Ganga River
(SP0)
6
Tur bid ity [N TU]
n
0
2
4
6
8
10
THM (SP3)
[µg/L]
THM (SP4)
[µg/L]
66
THM [µg/L]
n
0
1
2
3
4
5
6
Bank Filtrate
(SP1)
After AFM
(SP2)
After ECl2
(SP3)
Drinking Water
(SP4)
6666
n
b)
a) c)
Figure 13.
Turbidity removal (
a
,
b
) and THM formation (
c
) during short term test with optimized
system setting.
The second filter reduced the turbidity down to 0.55
±
0.08 NTU, after it had increased to
0.92
±
0.13 NTU behind the electrolytic cell, improving overall water quality. The THM analysis
showed concentrations of 2.2
±
0.5
µ
g/L in SP3 after 30 min and 4.2
±
2.1
µ
g/L in SP4. Those low
concentrations were expected due to the low TOC and dissolved organic carbon DOC content of the
water, allowing full compliance even with strict guideline values for DBPs (Table 1).
3.11. Energy Demand and Solar Energy Supply
During the first trial period a water volume of 1037 m
3
was treated and the total electricity
demand was summed up to 271 kWh without and 412 kWh with bank filtrate pumping. The average
flow rate through the system, including off times (e.g., at night, during maintenance or repair) in
Haridwar, was 180 L/h. This resulted in an average power demand of 46 W without and 70 W with
bank filtrate pumping, and a per m3energy consumption of 0.4 kWh (Table 5).
Table 5. Energy demand of the tested ECl2system.
Energy Demand
in [kWh/m3]
Water Pumping
through System incl.
AFM Filtration
Inline
Electro- lysis
Auxiliary and
Online Monitoring
Total Demand
Water Treatment
Pumping of Bank
Filtrate (BF)
Total Demand
incl. Pumping
RBF AFM ECl2
station in Haridwar
0.06 0.19 0.02 0.26 0.14 0.40
Water 2019,11, 122 14 of 17
The power requirement of 70 W was used for the sensitivity analysis using different sized
PV generators and different battery capacities considering the parameters mentioned in Table 3.
The analysis shows that at least a 96 Ah battery system with minimum 800 Wp are required to power
the water treatment system for 22 h per day and a permitted capacity shortage of 1%. The results are
summarized in Table 6.
Table 6. Results of sensitive analysis based on shown PV battery combinations.
PV Generator Size and
Battery Capacity [W], [Ah]
Total Prod.
[kWh/a]
Total Cons.
[kWh/a]
Input Battery
[kWh/a]
Output Battery
[kWh/a]
Excess Electricity
[kWh/a], [%]
Unmet Load
[kWh/a], [%]
Capacity Shortage
[kWh/a] [%]
600, 96 928 553 334 278 319, 34,4% 9.07, 1.6% 10.3, 1.8%
600, 144 928 560 343 285 310, 33,4% 1.90, 0.3% 2.13, 0.4%
800, 96 1237 559 332 277 622, 50.1% 3.10, 0.6% 3.51, 0.6%
800, 144 1237 561 335 279 619, 50.3% 0.80, 0.1% 0.89, 0.2%
900, 96* 1392 560 330 275 776, 55.8% 2.13, 0.4% 2.38, 0.4%
900, 144 1392 562 332 277 774, 55.6% 0.48, 0.1% 0.53, 0.1%
1000, 96 1546 561 339 274 930, 60.2% 1.58, 0.2% 1.78, 0.3%
1000, 144 1546 562 330 275 929, 60.1% 0.16, <0.1% 0.19, <0.1%
* installed solar PV-battery system combination.
The simulation shows, that with the installed solar PV-battery combination of 900 Wp and
2×96 Ah
, the required electricity can be supplied nearly throught the year. Only a few hours of
electrictiy shortages including a 7-h long power cut in the middle of the monsoon in August occured.
The total capacity shorage summed up to 0.4%. Whether this is permissible in a real case scenario and
whether this could be compensated by, for example, an increase of the water storage capacity depends
on local conditions. The simulated power production and SoC is shown in Figure 14a,b, respectively.
Water 2018, 10, x FOR PEER REVIEW 15 of 17
Figure 14. Power output from 900 Wp PV panel (a) and corresponding State of Charge (SoC) (b).
During the trial the station was continuously running on the given solar PV system, as long as
the modules were cleaned from dust frequently.
4. Conclusions
The presented results of a first long term and a second short term trial of a RBF ECl2
combination in India show that the tested system poses a feasible alternative for decentralized and
safe drinking water supply for river bound communities in developing countries. RBF serves as a
very efficient pre-treatment step, substantially reducing pathogens, turbidity, and DBP precursors.
The installed AFM filter is capable of further reducing the already low turbidity values and the ECl2
system completely removes all still-present indicator pathogens, and supplies sufficient residual
disinfectant for safe water distribution. The station complies with given water regulations
concerning indicator pathogen and chlorine concentrations. Additionally, the production of DBPs is
of no concern and stays well below the given guideline values. The first test period revealed some
optimization potential of the control algorithm and the system setting, which was successfully
implemented for the second trial. After that, the system reacted reliably to changing source water
and operating conditions by keeping the residual disinfectant at a constant level. The used ORP
sensor is able to indicate “sufficient” or “insufficient” disinfectant once it is given sufficient reaction
time. For more accurate and faster readings, the application of chlorine probes may be an alternative
to ORP probes.
Clogging of the electrolytic cell, due to increased levels of hardness, remains an operational
challenge of the ECl2, which needs to be specifically addressed. With polarity inversion alone, and
total hardness levels above 200 mg/L, the operation of an ECl2 system currently reduces the
maintenance intervals to about once every two months. The operation of the ECl2 system at total
hardness values above 300 mg/L is currently not advisable. With the implementation of an
additional probe to measure electrical conductivity, cell overgrowth could be detected by
monitoring current and voltage of the cell and comparing them with the actual conductivity of the
water.
After this trial, the system is mature enough to be implemented in a real scenario and under
favorable operational conditions and source water quality, and it should be possible to reduce the
maintenance intervals of the station to six months. For this, the implemented SCADA system will
play an important role. An increase of the treatment capacity is straight forward by increasing the
ECl2-cell size and the solar energy supply accordingly.
Author Contributions: P.O., A.G., and F.B. developed the pilot station; P.O. reviewed the literature, analyzed
the data, and prepared the draft of the publication; F.B. dimensioned the solar energy supply system; P.M.,
S.K.S., and P.C.K. took care of the infrastructure for the pilot study; P.M., G.N., and M.W. analyzed water
January April July October
January April July October
Hour of DayHour of Day
PV power output
8
00 W
0 W
4
00 W
100%
70%
40%
Batte ry s tate of charge
18
12
6
24
18
12
6
24
a)
b)
Reduced solar radiation during
monsoon caus es
(acceptable) unmet load
Figure 14. Power output from 900 Wp PV panel (a) and corresponding State of Charge (SoC) (b).
During the trial the station was continuously running on the given solar PV system, as long as the
modules were cleaned from dust frequently.
4. Conclusions
The presented results of a first long term and a second short term trial of a RBF ECl
2
combination
in India show that the tested system poses a feasible alternative for decentralized and safe drinking
water supply for river bound communities in developing countries. RBF serves as a very efficient
pre-treatment step, substantially reducing pathogens, turbidity, and DBP precursors. The installed
AFM filter is capable of further reducing the already low turbidity values and the ECl
2
system
completely removes all still-present indicator pathogens, and supplies sufficient residual disinfectant
for safe water distribution. The station complies with given water regulations concerning indicator
Water 2019,11, 122 15 of 17
pathogen and chlorine concentrations. Additionally, the production of DBPs is of no concern and stays
well below the given guideline values. The first test period revealed some optimization potential of
the control algorithm and the system setting, which was successfully implemented for the second
trial. After that, the system reacted reliably to changing source water and operating conditions by
keeping the residual disinfectant at a constant level. The used ORP sensor is able to indicate “sufficient”
or “insufficient” disinfectant once it is given sufficient reaction time. For more accurate and faster
readings, the application of chlorine probes may be an alternative to ORP probes.
Clogging of the electrolytic cell, due to increased levels of hardness, remains an operational
challenge of the ECl
2
, which needs to be specifically addressed. With polarity inversion alone, and total
hardness levels above 200 mg/L, the operation of an ECl
2
system currently reduces the maintenance
intervals to about once every two months. The operation of the ECl
2
system at total hardness values
above 300 mg/L is currently not advisable. With the implementation of an additional probe to measure
electrical conductivity, cell overgrowth could be detected by monitoring current and voltage of the cell
and comparing them with the actual conductivity of the water.
After this trial, the system is mature enough to be implemented in a real scenario and under
favorable operational conditions and source water quality, and it should be possible to reduce the
maintenance intervals of the station to six months. For this, the implemented SCADA system will play
an important role. An increase of the treatment capacity is straight forward by increasing the ECl
2
-cell
size and the solar energy supply accordingly.
Author Contributions:
P.O., A.G., and F.B. developed the pilot station; P.O. reviewed the literature, analyzed the
data, and prepared the draft of the publication; F.B. dimensioned the solar energy supply system; P.M., S.K.S.,
and P.C.K. took care of the infrastructure for the pilot study; P.M., G.N., and M.W. analyzed water samples
and contributed expertise in interpreting results; and C.S. and T.G. supported the planning and design of the
experiments and the preparation of the draft. All co-authors reviewed and edited the draft.
Funding:
All primary data was collected within the AquaNES project. This project has received funding from the
European Union’s Horizon 2020 Research and Innovation Program, under grant no. 689450
Acknowledgments:
This study was based on the excellent support from Uttarakhand Jal Sansthan (UJS),
which provided the pilot locations and supported in-system construction and operation, and the work performed
by many staff members of UJS. The authors also gratefully acknowledge support from S. Hertel, for supporting
literature review, and from F. Kowalczyk, F. Bauer, R. Rajoriya, F. Naumann, Binod Das, and P. Patwal during
sampling, water analysis, and system maintenance.
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
Onda, K.; LoBuglio, J.; Bartram, J. Global access to safe water: Accounting for water quality and the resulting
impact on MDG progress. Int. J. Environ. Res. Public Health 2012,9, 880–894. [CrossRef] [PubMed]
2.
Bain, R.; Cronk, R.; Hossain, R.; Bonjour, S.; Onda, K.; Wright, J.; Yang, H.; Slaymaker, T.; Hunter, P.;
Prüss-Ustün, A.; et al. Global assessment of exposure to faecal contamination through drinking water based
on a systematic review. Trop. Med. Int. Health 2014,19, 917–927. [CrossRef] [PubMed]
3.
Bain, R.; Cronk, R.; Wright, J.; Yang, H.; Slaymaker, T.; Bartram, J. Fecal contamination of drinking-water in
low- and middle-income countries: A systematic review and meta-analysis. PLoS Med.
2014
,11, e1001644.
[CrossRef] [PubMed]
4.
World Health Organization (WHO). Drinking Water Key Facts. 2018. Available online: http://www.who.
int/en/news-room/fact-sheets/detail/drinking-water (accessed on 2 October 2018).
5.
Hossain, M.A.; Sengupta, M.K.; Ahamed, S.; Rahman, M.M.; Mondal, D.; Lodh, D.; Das, B.; Nayak, B.;
Roy, B.K.; Mukherjee, A.; et al. Ineffectiveness and Poor Reliability of Arsenic Removal Plants in West Bengal,
India. Environ. Sci. Technol. 2005,39, 4300–4306. [CrossRef] [PubMed]
6.
Sobsey, M.D.; Handzel, T.; Venczel, L. Chlorination and safe storage of household drinking water in
developing countries to reduce waterborne disease. Water Sci. Technol. A J. Int. Assoc. Water Pollut. Res.
2003
,
47, 221–228. [CrossRef]
7.
Montgomery, M.A.; Elimelech, M. Water and Sanitation in Developing Countries: Including Health in the
Equation. Environ. Sci. Technol. 2007,41, 17–24. [CrossRef] [PubMed]
Water 2019,11, 122 16 of 17
8.
Roberts, L.; Chartier, Y.; Chartier, O.; Malenga, G.; Toole, M.; Rodka, H. Keeping clean water clean in
a Malawi refugee camp: A randomized intervention trial. Bull. World Health Organ.
2001
,79, 280–287.
[PubMed]
9.
Lorenzen, G.; Sprenger, C.; Taute, T.; Pekdeger, A.; Mittal, A.; Massmann, G. Assessment of the potential for
bank filtration in a water-stressed megacity (Delhi, India). Environ. Earth Sci.
2010
,61, 1419–1434. [CrossRef]
10.
Dash, R.R.; Bhanu Prakash, E.V.P.; Kumar, P.; Mehrotra, I.; Sandhu, C.; Grischek, T. River bank filtration in
Haridwar, India: Removal of turbidity, organics and bacteria. Hydrogeol. J. 2010,18, 973–983. [CrossRef]
11.
Weiss, W.J.; Bouwer, E.J.; Aboytes, R.; LeChevallier, M.W.; O’Melia, C.R.; Le, B.T.; Schwab, K.J. Riverbank
filtration for control of microorganisms: Results from field monitoring. Water Res.
2005
,39, 1990–2001.
[CrossRef]
12.
Sandhu, C.; Grischek, T.; Kumar, P.; Ray, C. Potential for Riverbank filtration in India. Clean Technol.
Environ. Policy 2011,13, 295–316. [CrossRef]
13.
Wang, J.; Smith, J.; Dooley, L. (Eds.) Evaluation of Riverbank Infiltration as a Process for Removing Particles and
DBP Precursors; American Water Works Association: Denver, CO, USA, 1996.
14.
Grischek, T.; Paufler, S. Prediction of Iron Release during Riverbank Filtration. Water
2017
,9, 317. [CrossRef]
15.
Sandhu, C.; Grischek, T. Riverbank filtration in India—Using ecosystem services to safeguard human health.
Water Sci. Technol. Water Supply 2012,12, 783–790. [CrossRef]
16.
Wintgens, T.; Nattorp, A.; Elango, L.; Asolekar, S.R. Natural Water Treatment Systems for Safe and Sustainable
Water Supply in the Indian Context: Saph Pani. Water Intell. Online 2016,15, 9781780408392. [CrossRef]
17.
Ghodeif, K.; Grischek, T.; Bartak, R.; Wahaab, R.; Herlitzius, J. Potential of river bank filtration (RBF) in
Egypt. Environ. Earth Sci. 2016,75, 255. [CrossRef]
18.
Pholkern, K.; Srisuk, K.; Grischek, T.; Soares, M.; Schäfer, S.; Archwichai, L.; Saraphirom, P.; Pavelic, P.;
Wirojanagud, W. Riverbed clogging experiments at potential river bank filtration sites along the Ping River,
Chiang Mai, Thailand. Environ. Earth Sci. 2015,73, 7699–7709. [CrossRef]
19.
Bartak, R.; Page, D.; Sandhu, C.; Grischek, T.; Saini, B.; Mehrotra, I.; Jain, C.K.; Ghosh, N.C. Application
of risk-based assessment and management to riverbank filtration sites in India. J. Water Health
2015
,13,
174–189. [CrossRef] [PubMed]
20.
Hashmi, I.; Farooq, S.; Qaiser, S. Chlorination and water quality monitoring within a public drinking water
supply in Rawalpindi Cantt (Westridge and Tench) area, Pakistan. Environ. Monit. Assess.
2009
,158, 393–403.
[CrossRef] [PubMed]
21.
Clasen, T.; Haller, L.; Walker, D.; Bartram, J.; Cairncross, S. Cost-effectiveness of water quality interventions
for preventing diarrhoeal disease in developing countries. J. Water Health
2007
,5, 599–608. [CrossRef]
[PubMed]
22.
Morris, R.D.; Audet, A.M.; Angelillo, I.F.; Chalmers, T.C.; Mosteller, F. Chlorination, chlorination by-products,
and cancer: A meta-analysis. Am. J. Public Health 1992,82, 955–963. [CrossRef]
23.
World Health Organization (WHO). Guidelines for Drinking-Water Quality; World Health Organization:
Geneva, Switzerland, 2017.
24.
Kraft, A. Electrochemical Water Disinfection: A Short Review. Platin. Met. Rev.
2008
,52, 177–185. [CrossRef]
25.
European Commission. Proposal for a Directive of the European Parliament and of the Council on the Quality of
Water Intended for Human Consumption (Recast); European Commission: Brussels, Belgium, 2018.
26.
Otter, P. Experimental Determination of the Optimization Potential of an Energy Autarkic Drinking Water
Purification System Under the Consideration of Local Conditions in the Rubber Trapper Reserve “RESEX do
Rio Ouro Preto” Located in the Brazilian State of Rondônia. Master’s Thesis, University of Kassel, Kassel,
Germany, 2010.
27.
Kraft, A.; Blaschke, M.; Kreysig, D.; Sandt, B.; Schröder, F.; Rennau, J. Electrochemical water disinfection.
Part II: Hypochlorite production from potable water, chlorine consumption and the problem of calcareous
deposits. J. Appl. Electrochem. 1999,29, 895–902. [CrossRef]
28.
Schmidt, W. Untersuchungen zur Desinfektionswirkung und Sicherheit der In-line-Elektrolyse von Chlor als
umweltschonendes Verfahren für die Desinfektion von Trinkwasser—In-line-Elektrolyse für die Trinkwasserdesinfektion
(Investigations on the Disinfecting Effectiviness and Safety of the In-Line Electrolysis of Chlorine as an Environmentally
Friendly Process for the Disinfection of Drinking Water In-Line Electrolysis for Drinking Water Disinfection); DBU
Report; Deutsche Bundesstiftung Umwelt: Osnabrück, Germany, 2012.
Water 2019,11, 122 17 of 17
29.
Haaken, D.; Dittmar, T.; Schmalz, V.; Worch, E. Influence of operating conditions and wastewater-specific
parameters on the electrochemical bulk disinfection of biologically treated sewage at boron-doped diamond
(BDD) electrodes. Desalin. Water Treat. 2012,46, 160–167. [CrossRef]
30.
Kraft, A.; Stadelmann, M.; Blaschke, M.; Kreysig, D.; Sandt, B.; Schröder, F.; Rennau, J. Electrochemical water
disinfection Part I: Hypochlorite production from very dilute chloride solutions. J. Appl. Electrochem.
1999
,
29, 859–866. [CrossRef]
31.
Deutscher Verein des Gas und Wasserfaches. DVGW W 296 (A): Trihalogenmethanbildung—Vermindern,
Vermeiden und Ermittlung des Bildungspotentials (Trihalogenmethane Formation—Reduction, Avoidance and
Determination of Formation Potential); Wirtschafts- und Verlagsgesellschaft Gas und Wasser mbH: Bonn,
Germany, 2014.
32.
Lilienthal, P.; Lambert, T.; Gilman, P. Homer: The Micropower Optimization Model; Homer Energy, LLC.:
Boulder, CO, USA, 2009.
©
2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
